The client is one of the world’s largest consumer goods companies and owns some of the best known global brands. Selling products in nearly 200 countries, their operational function is widespread and complex. In order to remain a leader in a competitive marketplace the company evolves constantly to meet changing business needs.
The client is developing a number of individual ‘product sites’ selling aspirational brands, direct to their consumers across various markets. This process is part of a wider B2C (Business to Consumer) initiative using a new strategic platform, Demandware.
The design and development process involves a number of parties; the end client, their marketing partners, branding/design agencies and the platform developers, who are responsible for both front-end website delivery as well as back-end system integration (e.g. order fulfilment and stock control).
The individual product sites were developed as discreet packaged launches across various countries, but ultimately they are all based on the same core platform and development process.
The client faced a challenge in ensuring thorough test coverage across the combination of brands within countries as well as ensuring web browser access compatibility. This challenge is particularly difficult for End to End (E2E) testing.
ROQ has a long relationship with the client through engagement on other projects so we were the natural partner to ensure this business change was implemented smoothly.
ROQ were already delivering a managed service for manual functional testing on these projects, and a key responsibility was to develop the manual approach in a way that made it easy to create an automated regression pack. The regression pack needed to cover the back and front end systems based on UAT user journeys.
Utilising their Automation Framework, ROQ were able to use the existing manual test scripts to form the basis of the automated regression pack. With the support of an automation expert, the Framework also enabled the manual testers to transition into building the automated scripts and as such, retain valuable system knowledge.
From the offset of the project, ROQ envisaged a role for automated regression testing. Neither the client nor the ROQ test team had the desire to manually test all brand sites, across all geographies, by language and then by browser and/or device.
ROQ took both a ‘top down’ and a ‘bottom up’ approach to automation. Our vision was that an E2E script would be decomposed into testable functions e.g. test Google+ login, test Facebook login, test product ’out of stock’, test product on back order, etc.
These units formed the basis of 60+ functional tests that verified key elements of the product site from the ‘bottom up’. Top down E2E regression tests were built based on these reusable units or functions.
ROQ executes all of these tests in a Continuous Integration (CI) cycle using Jenkins, so that tests run automatically and are synchronised with the daily releases into UAT.
Each day the regression tests run across 3 countries, namely the United Kingdom, United States and France.
Furthermore, for every country the test cases execute across three major devices, Desktop, Mobile and Tablet. Within these categories test are further classified by browser type IE, Chrome, Firefox or Safari.
In this flexible model, a single test can execute across 3 x countries x various devices x various browsers, resulting in multiple executions from one common base
ROQ provides test automation as a managed service. The client knew that they faced a real problem in regression testing across daily releases, on multiple platforms for each geography. ROQ took responsibility for managing the analysis, execution and reporting across this complex delivery cycle.
During the initial execution of this regression pack, ROQ highlighted a number of very significant showstopper defects that would have gone unrecognised from solely manual tests. This prevented major usability or functional defects reaching production.
The ROQ Framework is very adaptable, exhibiting the ability to transcend Desktop, Mobile and Tablet configurations. The solution is designed to minimise the dependency on technical experts who can drive test automation in a coding sense, and places more emphasis on ‘automation-savvy’ test analysts who can specify instructions to our automation engine.